AbstractSoybean (Glycine max (L.) Merrill) growers in Brazil commonly band corrective and maintenance fertilizer near the seed at planting time. Little information existed to substantiate this as the most efficient P placement method. Because P is an immobile element in the soil and a costly import for Brazilian farmers, placement may be criticalThis research was undertaken to evaluate the influence of P broadcast and incorporated vs. banded near the seed upon yield and leaf composition of field‐grown ‘Bragg’ soybean. The study was conducted on Sao Jeronomo silt loam (Paleudult), low in available P, at 30° S lat about 70 km west of Porto Alegre, Brazil. Experimental design the initial year was a 4 ✕ 4 factorial, three replications, randomized block with P‐rates of 0, 35, 70, and 140 Kg P/ha broadcast and/or banded as triple super hosphate. This established with the P broadcast and P incorporation, an available soil‐P gadient to determine at what levels banded‐P might most influence yield. Second year residual P responses was measured after thoroughly mixing the soil and measuring available P with Bray P2 (0.03N NH1F + 0.1N HCI) and Melhich (0.025N H2O4 + 0.05N HCl) extractors. Plant measurements include yield, lodging index and leaf N, P, and K composition. Statistical relationship between relative yield and Bray P2 available soil‐P were evaluated by linear, quatratic, second degree polynomial, logarithmic, and segmented lines methods. Correlation coefficients between P treatments, soil available P, leaf N, P, and K and lodging index provided additional interpretations. DRIS interpretation norms of leaf N, P, and K composition was used to evaluate the nutrient balance.The 1st year (1974) broadcast‐P was significantly correlated at the 1% level to eld (r = 0.41) and leaf‐P (r = 0.41) and at the 5% level to lodging index (r = 0.29). Banded‐P was significantly correlated at the 1% level to leaf‐P (r = 0.61), leaf‐K (r = 0.53), and lodging (r = 0.63), but yield, though increased, was not significant at the 5% level (seed yield ranged from zero P of 2,775 to 4,835 kg/ha for 140 kg P/ha broadcast compared to 3,936 k e a for this quantity banded). The maximum lodging index for broadcast P only was 2.7 (about 17% lodged) while banded‐P only increased lodging index to 4.3 (about 46% lodging). This suggested an effect of high P levels upon increased lodging that was more pronounced by banding the P resulting in lower yields. Mixing the soil before the 2nd year incorporated the previously banded‐P so that the residual P responses were quite similar to the broadcasting of P. Leaf‐P was correlated significantly at the 1% level from both broadcast‐ P (r = 0.49) and previously banded‐P (r = 0.57) and yields correlated at the 5% level (r = 0.31 for broadcast‐P and r = 0.36 for previously banded‐P). The seed yield ranged from zero P of 1,945 to 3,150 kg/ha for 140 kg P/ha compared to 3,338 k o a to this quantity previously banded.The Bray P2 extractor correlation was higher with broadcast‐P and leaf‐P than with Melhich extractor and, therefore, was used to interpret relationship between relative yield and available soil‐P. The segmented line model attributed 32% of relative yield to Bray P2 available soil‐P, which was superior to the linear, quatratic, second degree polynomial, and logarithmic. This model indicated the maximum (93.2%, relative yield) was obtained at 8.83 ppm‐P, with no increase in relative yield with increasing levels of Bray P2 available‐P.DRIS evaluation of leaf N, P, and K composition suggest that for zero P treatments the order of nutrient needs were P > N > K for both years of the study. In the 1st year, the addition of P changed the order of need from N > P > K. The 2nd year, the order of nutrient need was P > N > K, regardless of soilavailable P level, and even when leaf‐P levels would not be classified as low or deficient. Leaf‐N and ‐P varied more during the 2 years of the study than did leaf‐K. High leaf‐K levels suggest this nutrient was high with respect to N and P.Results from this study suggest that banding P near the seed increased P uptake vs. broadcast and incorporation of P, but did not provide highest yields. The plantsoil growth factors obtained from broadcasting P resulted in highest return from P inputs.